XBRL 1st Annual Academic Competition

Participant's Packet

The Making of a Taxonomy

Text Box:


 

by Eric E. Cohen, CPA, PricewaterhouseCoopers LLP

Eric.E.Cohen@us.pwcglobal.com  eric.e.cohen@us.pwcglobal.com

XBRL 1st Annual Academic Competition

The Making of a Taxonomy............................................................................................................

Before you get started.............................................................................................................................

So your first question is: What's a taxonomy?

Now your second question may be: "How do I go about creating a taxonomy?"...................................

Nuts and bolts: Steps to create a taxonomy

1. Consider an important aspect of business reporting that can benefit from being expressed in XML, to enable cross-platform, application and operating system independent sharing of this information between applications, systems, and reporting entities.     

2. a) Assemble a group of knowledge experts, and/or b) Find resources with examples that illustrate the necessary output to begin determining the terms of your vocabulary...............................................................................................

3. Determine underlying authority or general practice for representing that information.......................

4. Consider where that information will be coming from, in particular if the input is from automated or manual systems.

5. Consider how to best organize the information...................................................................................

6. Consider how this taxonomy relates to the existing C&I taxonomy for reuse and comparability

7. Come up with the best way to represent the concepts as terms

Identify the concepts...........................................................................................................................

Determine the best shorthand for representing that concept..............................................................

8. Capture the terms in an electronic format: XML, an Excel or Lotus spreadsheet or a database.........

9. Document, document, document.........................................................................................................

Before you get started

So you want to create an XBRL taxonomy! Good for you. You have started down an important path.

So your first question is: What's a taxonomy?

Fair question. A taxonomy is a vocabulary, the foundation of sharing XML documents. It is the core that lets applications and organizations share information quickly, easily, and inexpensively. It is the structure and nomenclature for its field. It is the authoritative lookup file, distributed between users of shared information, to which references point and from which copies and updates flow. It is a classification system, a systematic representation of a body of knowledge. Why is that important?

If you spend time finding things on the Internet using search engines, you know that there are a number of tools. Each has strengths and weaknesses. An engine like Yahoo (www.yahoo.com) is great at context, because people choose where on Yahoo's hierarchical tree their site should be listed. The entries in Yahoo's hierarchical tree are like a taxonomy. Yahoo does not have to test to make sure the site belongs in that exact category. Close is good enough.

However, Yahoo does not list as many sites as Altavista (www.raging.com) or Northern Light (www.northernlight.com). These sites send computer agents onto the Web to create huge indexes of words. However, there is no ability to search for those sites in context.

If a site could use special tags (XML) within a document to provide the context of that document, and search engines knew how to look for those tags, you could get the best of a contextual engine like Yahoo and a content engine like Altavista. All you would need is a consistent set of tags.

People are willing to put their site into Yahoo's hierarchical category tree, even if it involves compromise, so embedding a pre-developed set of tags into documents will provide benefits, even if they are not perfect. In our world, this set of tags describes our documents and reports. The words and relationships expressed by these tags is our taxonomy.

Our taxonomy provides:

         A unique identifier for each concept we are working with

         The generally used, or common, English name of the item

         The immediate parent for that concept

         The mathematical relationship of the child to the parent

         A description of the item

         A customary order that the item is presented in a financial statement

         The underlying authority or documentation that shows why the item is included in the taxonomy

 

Now your second question may be: "How do I go about creating a taxonomy?"

If so, we can help you, and lead you through the nuts and blots of creating a taxonomy.

Text Box:


 


If your second question was "What time is lunch?" or "What do we win?" we suggest you try alternative resources. If it is "What's XBRL?" we will direct you to introductory information in your participant's packet to learn about the Extensible Business Reporting Language (www.xbrl.org).

Nuts and bolts: Steps to create a taxonomy

Text Box: Steps to create taxonomy:
1.	Consider an important aspect of business reporting that can benefit from being expressed in XML, to enable cross-platform, application and operating system independent sharing of this information between applications, systems, and reporting entities.
2.	Find resources with examples that illustrate the necessary output to begin determining the terms of your vocabulary.
3.	Determine underlying authority or general practice for representing that information
4.	Consider where that information will be coming from, in particular if the input is from automated or manual systems.
5.	Consider how to best organize the information - how that information is related in a hierarchy; determine levels of detail and necessary summarization. Most data can be presented differently depending on your perspective.
6.	Consider how this taxonomy relates to the existing C&I taxonomy for reuse and comparability.
7.	Come up with the best way to represent the concepts as terms.
8.	Capture the terms in an electronic format: XML, an Excel or Lotus spreadsheet or a database.
9.	Document, document, document.


 


1. Consider an important aspect of business reporting that can benefit from being expressed in XML, to enable cross-platform, application and operating system independent sharing of this information between applications, systems, and reporting entities.

There are numerous current studies that discuss the need for the electronization of financial information.

These include:

        The Impact of the Internet on the Future of Financial Reporting, Stephen M.H. Wallman, The Brookings Institution

         Business Reporting Research Project; Electronic Distribution of Business Reporting Information, FASB

         Business Reporting on the Internet, Lymer, DeBreceny, Gray and Rahman

         The Potential  Implications of Electronic Online Reporting on Financial and Business Reporting, Gerard Trites, FCA

 

We chose to start off with the representation of a financial statement. We also considered the representation of general journal transactions as a first step and are working on that with the first taxonomy and technical specification as guides.

 

As we consider the direction of XBRL, there are many areas in the business reporting chain that need to be incorporated into the XBRL framework. These include tax filings, audit workpapers, SEC filings, and emerging business metrics. These can all be tied into the basic technology and share information with the first taxonomy.

2. a) Assemble a group of knowledge experts, and/or b) Find resources with examples that illustrate the necessary output to begin determining the terms of your vocabulary.

 

Once the area of concentration has been determined, a professional group can seek out a team of knowledge, or domain, experts with education, experience and skills in the area up for consideration. For a group of students, this is not as easy to do. Seeking out "bottled brains" may be the next best choice.

 

As an organization that started its life within the American Institute of Certified Public Accountants, XBRL.org's members thought they knew a thing or two about the creation of financial statements. They did. But each firm has its own ways of doing things. Getting so many different firms around the table to talk about the contents of a financial statement was an interesting process.

 

For financial statements, there are many resources for finding illustrative disclosures. One method might have been picking a random sample of companies and reviewing their physical financial statements from the library, reviewing their filings at the SEC EDGAR site (www.sec.gov/edgarhp.htm) or downloading HTML or PDF files from their corporate investor web pages.

 

XBRL.org approached this task by using an AICPA document published annually, called Accounting Trends & Techniques.  This volume is an annual survey of accounting practices followed in 600 Stockholders' Reports. Our task force reviewed the guide to find general practices. They then compared those results with the guides and checklists their individual firms use in preparing financial statements.

 


 

3. Determine underlying authority or general practice for representing that information.

 

In taxonomy development, as in business practice, there are two kinds of standards: de jure and de facto. A de jure standard is one that is published by a standards-setting group. A de facto standard is a practice that is so popular it is generally accepted as a rule with as much authority as a de jure standard. No one ever required software developers to pop up a help key when you press F1. However, we expect that to be the case.

 

When you develop a taxonomy, there has to be an underlying reason to include a vocabulary term in your taxonomy. In the case of the financial statement vocabulary, we came up with over 1800 concepts and know that we could have included many more. Our goal was to come up with the concepts that are absolutely necessary for financial reporting, and to add to those terms only those optional concepts that are so much a part of everyday reporting that they are accepted as standards within an industry or the profession.

 


 

4. Consider where that information will be coming from, in particular if the input is from automated or manual systems.

 

Where will the data that comes together to populate your reports or databases come from? A number of important decisions will focus on this.

 

If all of your data comes from an extraction from one database, you will have fewer decisions to make than if data is:

         Built by assembling information from multiple data sources

         Combined from multiple sources that are similar but not designed uniformly, or

         Created by bringing together information that is created manually and trying to combine it with other sources.

Financial reporting involves [accounting information found in most systems] + [textual data found in text warehouses or in word processing files] + [manual tables created in spreadsheets].

General journal transactions involve the combination of elements common to every accounting system (account, amount, date) with a myriad of other elements - everything from summaries of invoicing customers and paying vendors and employees to job costing detail.

5. Consider how to best organize the information

Think hierarchically. XML documents contain information that conforms to a tree structure. Information is inter-related. It is your job to describe how.

When you work with Microsoft Windows, and need to find your way around the resources available to you, you are working with a hierarchy. "My Desktop" is as high as it goes. Then you see "My Computer", "Network Neighborhood", "Recycle Bin" and folders directly on the desktop. "My Computer" has as children the drives available to you, the "Control Panel" and other configuration "folders". You can dig into a local or network drive and find folders, sub-folders, sub-sub-folders, and the files scattered throughout.

Now think of your project in that same way.

Text Box:

 

The structure of information - how that information is related in a hierarchy - can be as important as the vocabulary itself. For a printed financial statement, determining the structure was relatively easy - look in the table of contents, and you will see the breakdown ... with a few extra categories thrown in for good measure.

XML is about parent/child relationships. XML rolls up to one parent of all parents - called the root. How deep do you go? The intended reporting and querying will help determine the levels of detail and necessary summarization you will build in.

Why isn't this always easy? Most data can be presented differently depending on your perspective. You might think of these different perspectives by the answer to the question, "How would you like that sorted?"

For general ledger entries, they may be sorted by date, by account, by source journal and be grouped by department, branch or posting status.

 

6. Consider how this taxonomy relates to the existing C&I taxonomy for reuse and comparability

The first XBRL deliverable, the US GAAP Commercial and Industrial Financial Statement, was developed over a period of nine months by a tremendously knowledgeable group of accounting and technology experts. The concepts contained within that taxonomy have parallels between industry sectors, across jurisdictions, and even between reporting types.

Although we expect that this first taxonomy will evolve to incorporate benefits found in future taxonomies, it is important to place great value in tying in new taxonomies to the existing taxonomy.

An international taxonomy shares many concepts with the US GAAP taxonomy. Being able to map between a similar concept of cash or net income is important, and brings extra value to both taxonomies. A General Journal taxonomy needs to relate to the financial reporting taxonomies. If enough is not kept consistent between the US GAAP taxonomy and that of other industry sectors or jurisdictions, the value of the specification for the alternate reporting type is diminished.

7. Come up with the best way to represent the concepts as terms.

When the members of the group that has turned into XBRL.org got started with the taxonomy for Commercial and Industrial financial statements presented according to US GAAP (generally accepted accounting principles), they had an end in mind. That goal was to provide a name for every number, major block of text, and concept contained within a financial statement. That required the group to move beyond what is obviously visible in the document's structure to thinking how people may wish to access the information contain in that document.

Identify the concepts

One way to think through this perspective is to pretend that you are on the phone. You are trying to communicate the information found in the financial statement to someone on the other end of the phone.

You: "I have a financial statement in my hand."

Them: "What company is it? For what years? Who was the auditor? Did it receive a clean opinion?"

The first two questions are pretty easy. That information is usually sitting on the cover, and at the top of most of the major statements. The next questions are a little harder. The information is contained in the Accountants' Letter, and the auditor's name is normally found at the bottom ... but determining if it was a clean opinion takes some interpretation.

You need an element for "entity", "period", "auditor name" and "opinion type" ... and determining the opinion type requires adding information to the basic information.

Them: "What was their net income this year? How does it compare with last year?"

These questions may seem easy, but when you answer the question, you will be doing more than looking up a number at the intersection of "net income" and "this year" or "last year". You will be looking for clues like currency and implied zeroes (expressed in thousands,  millions, or billions).

You need elements for "net income" at a point in time, "currency", and "implied zeroes".

If information is embedded within a block of text, it needs to be tagged somehow. The text as a whole itself is important; so are important facts embedded within that text.

Determine the best shorthand for representing that concept

Having identified the basic concepts, you need to come up with a unique identifier for each concept. It must be short enough so that someone typing it twice will not have dire thoughts about your future.

It can be descriptive, so that a human interpreter can more easily identify it ... if that human identifier happens to speak the same language as you do. Then you can use standard or generally accepted terms and abbreviations to identify each concept.

It could be also be numeric, with no login in it whatsoever - harder to work with for you or a same-language interpreter, but not subject to language limitations. In addition, if you use a mnemonic (memory device or shorthand) and the terms change - such as the change from the nomenclature "Sub S" to an "S Corporation" in taxes - your mnemonic is more a liability than an asset. And the word "table" can mean so many things in so many different situations ... a number removes the confusion buy eliminating the issue.

8. Capture the terms in an electronic format: XML, an Excel or Lotus spreadsheet or a database.

The ultimate goal for the taxonomy is to be expressed electronically is so it can be shared and so documents that use the terms can be validated against a master list of the terms. It will eventually be turned into XML.

 

The tools for working with XML are still rudimentary. Few people have expertise for working directly with XML files. There are a number of XML editing tools - one we recommend is called XML Spy (www.xmlspy.com) - but for working on developing this vocabulary, more traditional tools like spreadsheets and databases, may be practical.

 

9. Document, document, document

Documentation of the steps you take in developing your taxonomy is critical. Decisions that seem so obvious today can be forgotten tomorrow, and the benefits of the original decision lost. Document your reasons for decision making. Document your process

Taxonomies are living and breathing. Underlying rules and practices evolve and change. That means that a developed taxonomy will undergo change. As new entries are added, old ones are modified or removed, or the structure must be completely rethought, the documentation will be vitally important for making the right choices.

 

Back to Academic Competition Web Page XBRL.org Main Page
XBRL Educational Resource Center At Bryant College Email Questions Here